Embodiments of the present invention provide an approach to mitigate the risk of downtime or failed upgrades to a networked computing environment (e.g., a cloud computing infrastructure). This approach generally comprises pre-testing and evaluating potential upgrade paths to identify potential problems, and optionally recommending one or more optimal configurations. Specifically, if a proposed upgrade is identified as having potential points of failure (e.g., based on pre-testing), an end user/customer may be so notified. If the upgrade has several potential successful configurations, such configurations may be evaluated to identify a configuration that appears to have the greatest likelihood of a successful installation scenario. Optionally, the end user may then choose to redirect workload(s) to a successfully configured system that has been provisioned and pre-tested.